计算机工程与应用2009，Vol.45Issue(33)：31-34,4.DOI:10.3778/j.issn.1002-8331.2009.33.011
改进GP分形理论的最近邻序列预测算法
Nearest neighbor series predicate algorithm based on improved GP fractal theory
摘要
Abstract
This paper analyzes the disadvantage that the subjectivity is too strong in existing time serials and predicate method. The fraetal theory is used for the time series prediction.The computing of GP algorithm and multiple autocorrelation algorithm are improved,and the reconstruction of the phase space is easier.After that the nearest neighbor of accumulation sampling path is se-lected for one time predicate in phase space.The new algorithm is more suited to reconstruction and predicting in the phase space.By validating at two time series dataset,the analysis resuh of this method is steady and exact,predication precision of it is high and the running time is short.关键词
时间序列/分形/GP算法/复自相关/最近邻预测Key words
time series/fractal/Genetic Programming(GP) algorithm/multiple autocorrelation/nearest neighbor predicate分类
信息技术与安全科学引用本文复制引用
吕威,马维曼,毕全成,黄健聪..改进GP分形理论的最近邻序列预测算法[J].计算机工程与应用,2009,45(33):31-34,4.基金项目
国家自然科学基金(the National Natural Science Foundation of China under Grant No.10171113.No.10471156). （the National Natural Science Foundation of China under Grant No.10171113.No.10471156）
